EIPsInsight

Loading Experience

Preparing your insights...

Loading0%
SSZ2026-07-10_002

SSZ Engine API #002

2026-07-10 2 decisions 459 transcript lines

Transcript

Ask about this call

Answered from this call's summary and transcript only.

AI-generated — verify against the recording.

Call summary

Targets
  • Next SSZ Engine API call in 2 weeks (approx. July 24, 2026); top-up sync and namespace scoping documents to be circulated beforehand — 00:51:29
Decisions
  • Agreed SSZ Engine API must specify which ETH-namespace endpoints are available on the engine port, consistent with Beacon API encoding conventions — 00:43:54
  • ETH-namespace method list for SSZ Engine API should be as minimal as possible; no automatic carry-over from JSON-RPC engine spec — 00:49:50
Highlights
  • Testing Updates:
    • ·SSZ support from project ZKEVM being integrated into EELS Amsterdam branch; reference vectors and REST abstraction to follow — 00:05:05
    • ·EELS planned to eventually serve engine requests as a temporary EL for CL testing; FastAPI/OpenRPC targeted short-term — 00:05:36
  • Ssz Engine Api Design:
    • ·Jacek (Nimbus) proposes REST path alignment with Beacon API namespace for coexistence on same port; especially relevant for unified clients — 00:07:46
    • ·Top-up sync gap identified: current Engine API has no reliable way for a CL to determine where to begin backfilling to an out-of-sync EL head — 00:08:56
    • ·Proposal: expose ETH get-latest-canonical-block (FCU-marked head) on engine port to enable CL-driven top-up sync without EL p2p — 00:36:30
    • ·SSZ REST Engine API currently lacks equivalents for ETH-namespace methods (chainID, syncing, etc.) that existing JSON-RPC engine spec guarantees; needs explicit scoping — 00:43:54
Action Items
  • Marius Van Der Wijden — Prepare document explaining current ETH-namespace method list on engine port, rationale, and candidates for removal; bring to ACDE and ACDC for broader input — 00:49:11
  • Tamaghna Chourdhuri / Dustin — Publish public document on top-up sync design (incorporating Jacek's HackMD and chain-following mechanism) to gather input from absent clients — 00:50:08

Meeting chat35

  • Soubhik

    gm

  • Justin Traglia

    Gm

  • Tamaghna Chourdhuri

    Gmgm!

  • Tamaghna Chourdhuri

    Lets start in 5 min

  • Jun Song

    gm

  • Sameer Agarwal

    gm

  • Łukasz Rozmej

    and me

  • Marius Van Der Wijden (M)

    And you :D

  • Dustin

    yes, sure (re Nimbus/Jacek comments)

  • Tamaghna Chourdhuri

    https://github.com/ethereum/pm/issues/2145#issuecomment-4934505945 https://hackmd.io/siEIwl9QTlODSZ6DYH3iqg

  • Marius Van Der Wijden (M)

    How would 1 port work for non-unified clients?

  • FLCL

    1 port for both jsonrpc + rest?

  • FLCL

    Two different services, not sure if need to be united

  • FLCL

    A reverse proxy thing can do this

  • FLCL

    It sounds like we just should not prohibit it and that's it

  • Dustin

    https://discord.com/channels/595666850260713488/1295427781722898505/1525091128867557427

  • FLCL

    Move p2p and tx gossip from el to cl?

  • Łukasz Rozmej

    who is the child of who again? ;)

  • FLCL

    we need to store meta about most recent beacon inserted block with state and return it to you?

  • FLCL

    but you can also hold it on cl side?

  • FLCL

    unless you have db reset

  • FLCL

    what if you are on diff chain

  • FLCL

    how pivot can help if we have no state for it?

  • Marius Van Der Wijden (M)

    "We previously got an fcu for it"

  • FLCL

    final block sounds safer, maybe less efficient

  • FLCL

    maybe both head+final in case of having state

  • FLCL

    if we restify get block, we can return blocks/head, blocks/final or so

  • Dustin

    https://github.com/ethereum/execution-apis/blob/60dbef739bde0db7ac30b6075439e192bfbd087e/src/engine/common.md#underlying-protocol

  • Dustin

    Jacek wrote a hackmd on it

  • Dustin

    on "top-up sync"

  • Tamaghna Chourdhuri

    https://hackmd.io/siEIwl9QTlODSZ6DYH3iqg This one right:)

  • Dustin

    yeah

  • Marius Van Der Wijden (M)

    Early call lets goo

  • Justin Traglia

    Reacted to "Early call lets goo" with 👍

  • Tamaghna Chourdhuri

    Lmao And I was thinking this might end in half an hour

Key decisions

  • Agreed SSZ Engine API must specify which ETH-namespace endpoints are available on the engine port, consistent with Beacon API encoding conventions

  • ETH-namespace method list for SSZ Engine API should be as minimal as possible; no automatic carry-over from JSON-RPC engine spec